/irc-logs / w3c / #webapps / 2009-02-02 / end
Options:
- # Session Start: Mon Feb 02 00:00:01 2009
- # Session Ident: #webapps
- # [00:09] * Joins: heycam (cam@130.194.72.84)
- # [01:53] * Quits: heycam (cam@130.194.72.84) (Quit: bye)
- # [02:09] * Joins: gavin__ (gavin@63.245.208.169)
- # [02:09] * Quits: gavin (gavin@63.245.208.169) (Connection reset by peer)
- # [03:28] * Joins: heycam (cam@130.194.72.84)
- # [04:47] * gavin__ is now known as gavin
- # [05:43] * Quits: shepazu (schepers@128.30.52.30) (Client exited)
- # [05:44] * Joins: shepazu (schepers@128.30.52.30)
- # [06:32] * Quits: heycam (cam@130.194.72.84) (Quit: bye)
- # [07:17] * Joins: heycam (cam@124.168.33.158)
- # [09:03] * Joins: tlr (tlr@128.30.52.30)
- # [09:28] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [10:25] * Joins: arve (arve@213.236.208.22)
- # [10:47] * Joins: anne (annevk@213.236.208.22)
- # [10:52] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: This computer has gone to sleep)
- # [11:05] * Quits: arve (arve@213.236.208.22) (Client exited)
- # [11:06] * Joins: Lachy (Lachlan@213.236.208.22)
- # [11:07] * Joins: arve (arve@213.236.208.22)
- # [13:02] * Joins: ArtB (ce846302@128.30.52.43)
- # [13:08] <marcos> Arve, I wondering what happened to using HTML5 Storage interface in the API spec?
- # [13:29] <arve> marcos: I ripped it out, and reintroduced our own on purpose, given that it's fairly impossible to refer to HTML5 normatively
- # [13:29] <arve> (There are also implementation concerns)
- # [13:31] <marcos> ok. Anil form MS also had concerns about storage because of it's dependency on origin.
- # [13:31] <anne> why is it impossible? several specs do it
- # [13:32] <marcos> Anne, which specs?
- # [13:32] <anne> you can just say that a widget has its own unique origin, you prolly need to say something about that anyway to interact with HTML features
- # [13:32] <anne> XHR, Web Workers, several of the more recently splitted out parts of HTML5
- # [13:32] <marcos> yeah, that's what I was thinking also
- # [13:33] <marcos> origin just needs to be some URI right?
- # [13:34] * arve recognizes «some URI» as contentious
- # [13:34] * marcos tries to put the square widget:// or tag:// square peg into the origin hole :D
- # [13:37] <anne> it's either a scheme, hostname, port tuple or an identifier of some sorts
- # [13:42] <arve> marcos: that's what I was getting at
- # [13:44] <anne> arve, you need to resolve the issue somehow, otherwise e.g. URI resolving and <iframe> etc. will fail
- # [13:44] <anne> and scripting won't work, etc.
- # [13:44] <anne> origin is not just needed for localStorage
- # [13:44] <anne> most stuff depends on it
- # [13:48] * Quits: arve (arve@213.236.208.22) (Quit: Leaving)
- # [13:49] * Joins: arve (arve@213.236.208.22)
- # [13:52] * Quits: arve (arve@213.236.208.22) (Client exited)
- # [13:52] <marcos> amen
- # [13:58] * Joins: arve (arve@213.236.208.22)
- # [14:04] <ArtB> marcos, yt?
- # [14:04] <marcos> ye
- # [14:04] <marcos> p
- # [14:05] <ArtB> the API & Events spec defines generic {get,set}Preferences APIs so the user agent can get/set all of the metadata in the config file, rigth?
- # [14:07] <marcos> no
- # [14:07] <ArtB> or is the idea the conf file's predefined metadata is retrieved from the attributes?
- # [14:07] <marcos> yes
- # [14:07] <marcos> Though, option 1 would have been more extensible
- # [14:08] <marcos> option 1 being get/set as you mentioned
- # [14:08] <ArtB> If someone wants to extend the config file, e.g. to add their own key+value pairs, how is that done?
- # [14:08] <ArtB> define their own element in their own NS?
- # [14:09] <marcos> through a namespace. however, then there is no mechanism for accessing those values at runtime
- # [14:10] <ArtB> Perhaps there is a need for app-specicif or Widget-specific name+value support in the config file
- # [14:10] <ArtB> <feature> seems kinda close
- # [14:10] <marcos> you mean for setting preferences?
- # [14:11] <ArtB> any of the widget's metadata
- # [14:12] <ArtB> I'm thinking about widget-specific settings
- # [14:12] <marcos> hmmm.... wonder if you could use xhr to get you the config document and then use getElementsByTagNameNS?
- # [14:12] <ArtB> how does the author state those?
- # [14:13] <ArtB> need something in the config file - config does mean configuration i.e. settings
- # [14:13] <marcos> sorry, I'm a bit lost :( Lets take a step back. What do you mean by widget-specific settings?
- # [14:13] <ArtB> <setting name="some_name" value=
- # [14:14] <ArtB> oops, let me try again
- # [14:14] <ArtB> <setting name="some_name" value="some_value"/>
- # [14:14] <marcos> these settings, would they be used at runtime by the start file?
- # [14:14] <ArtB> then getPreference("some_name") returns "some_value"
- # [14:15] <ArtB> they would be used by the widget engine
- # [14:15] <marcos> ok, the way I was thinking that would work was like this:
- # [14:15] <ArtB> The current mode appears to this:
- # [14:15] <ArtB> <vendor:setting name="some_name" value="some_value"/>
- # [14:18] <marcos> In javascript, the author would check if a preference exists.
- # [14:18] <marcos> <html>
- # [14:18] <marcos> <script>
- # [14:18] <marcos> /check if prefs have been initialized
- # [14:18] <marcos> if(!preferences.getPreference('somePref')){
- # [14:18] <marcos> preferences.setPreference('somePre','someValue');
- # [14:18] <marcos> preferences.setPreference('somePre2','someValue2');
- # [14:18] <marcos> etc.
- # [14:18] <marcos> }
- # [14:18] <marcos> </script>
- # [14:19] <marcos> my rationale was to keep code where the code is, and away from the configuration side of things.
- # [14:22] <marcos> From my point of view, this is less annoying for developers as it means that they only set preferences in one place
- # [14:23] <ArtB> I see your point. But OTOH, having the settings/prefs/config-data in the declarative format make sense too
- # [14:24] <marcos> it does, I agree. But I don't see any advantage.
- # [14:30] <marcos> Having things that are used in code (preferences) being declare in the configuration document still seems like a big pain to me. It just replicates functionality.
- # [14:31] <ArtB> YA example of more than one way to "skin a cat" :)
- # [14:32] <ArtB> btw, have you received any requests to extend the LCWD comment period?
- # [14:32] <marcos> no, I have not
- # [14:34] <marcos> I'm pretty happy with the current reviews. However, it would be good if we could get an Anne or a Lachy to go through it and rip it apart (hint hint! :) )
- # [14:35] <Lachy> marcos, which spec?
- # [14:35] <marcos> Lachy: http://dev.w3.org/2006/waf/widgets/
- # [14:36] <Lachy> marcos, mail me the link to it and I'll take a look some time this week
- # [14:36] <Lachy> but also feel free to ping me daily until its done
- # [14:36] <marcos> Awesome! thanks Lachy.
- # [14:45] <marcos> fwiw, Artb, Danbri also raised similar issues at some point about having an extensibility point in the metadata. He even put together a proposal as to how GRRDL could be used.
- # [14:51] <ArtB> Marcos, are you still tweaking A&E spec?
- # [14:51] <marcos> nope
- # [14:51] <ArtB> ... in advance of FPWD, that is
- # [14:51] <ArtB> Arve, has the ball now?
- # [14:51] <marcos> yep
- # [14:54] <ArtB> Re "Objects implementing [HTML5]'s Window interface MUST also implement the WindowWidget interface. The WindowWidget interface provides access to [DOM3Core] Document objects in a widget context."
- # [14:54] <ArtB> ... should WindowWidget just be Widget?
- # [14:55] * Quits: arve (arve@213.236.208.22) (Client exited)
- # [14:55] <ArtB> oops; nevermind
- # [15:01] * Joins: arve (arve@213.236.208.22)
- # [15:06] <marcos> I'm worried about the issues you raised about the dependencies on specs. I'm not sure I understand the problem completely.
- # [15:06] <marcos> (above directed at Artb)
- # [15:06] <ArtB> A prob we may have is that P&C was at one time the 1 and only 1 spec
- # [15:07] <ArtB> It uses WUA (Widget User Agent) quite broadly
- # [15:07] <marcos> yeah
- # [15:07] <ArtB> P&C today is just about the manifest and package
- # [15:08] <ArtB> so if P&C normatively prescribes a WUA, that probably isn't correct
- # [15:09] <marcos> ok, I see what you mean
- # [15:09] <ArtB> I suspect cleaning up that terminology may be a bit painful
- # [15:09] <ArtB> so I don't think it must be done
- # [15:09] <ArtB> However, it will continue to create some confusion
- # [15:09] <marcos> it could be ok, but I'm still worried that a WUA is not defined (or it is defined, but just not explicitly).
- # [15:09] <ArtB> The P&C spec is NOT the right place to talk about a broad WUA, IMHO unless that "talk" is non-normative
- # [15:10] <marcos> ok, it should talk about a user agent that processes packages.
- # [15:10] <ArtB> bingo!
- # [15:10] <marcos> right. Ok. I'm with ya now.
- # [15:11] <marcos> This leads to the next issue
- # [15:11] <marcos> the problem of the dependency on the dig sig spec.
- # [15:12] <ArtB> I can see that DigSig will have a dependency on P&C e.g. for definitions
- # [15:13] <marcos> I don't see that dependency being different than the dependency on DOM3Core. We need DOM3Core, for instance, to give us TextNodes. As we need the Dig Sig spec to give us a status for the validity of a certificate.
- # [15:13] <ArtB> But you're saying it goes the other way?
- # [15:14] <marcos> We need to set a variable in the P&C spec based on how the user agent in the dig sig spec processes the signatures. The Dig Sig user agent's sole purpose is to provide that information to the packaging agent (or to other user agents as they may need to revalidate at runtime).
- # [15:15] <marcos> However, it could be the other way around
- # [15:16] <marcos> I.e., the widget dig sig spec could say something like "create variable X in the packaging specification and set it to be true". That seems wrong to me, however.
- # [15:18] * Quits: arve (arve@213.236.208.22) (Quit: Leaving)
- # [15:43] <marcos> Artb, I changed the packaging spec with regard to having a user agent that only operates on the package.
- # [15:46] <marcos> Responding to your email about this issue and outlining the changes
- # [15:51] <ArtB> Marcos, sigh on the DigSig dependency
- # [15:51] <marcos> I think we need it
- # [15:51] <marcos> But it's optional
- # [15:51] <ArtB> The change you made in the intro to Step #4 is the main thing I wanted
- # [15:51] <marcos> it was a totally valid point to make it optional.
- # [15:52] <ArtB> I'd love to say "and btw, if you support Wid-DigSig go there" and be done with it
- # [15:52] <ArtB> but maybe that's not possible ...
- # [15:53] <ArtB> Like you said, we need to hear from FH and MP and TLR and ...
- # [15:54] <ArtB> I think MP had a related point in http://lists.w3.org/Archives/Public/public-webapps/2009JanMar/0255.html
- # [15:57] <ArtB> It kinda' feels like we may not have the right work split re clear definition of the policy (e.g. what is valid), the policy decision point and the policy enforcement point
- # [15:59] <marcos> agreed
- # [16:20] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [17:02] * Quits: gsnedders (gsnedders@86.136.52.180) (Quit: Killin' teh intarwebs)
- # [17:03] * Joins: gsnedders (gsnedders@86.136.52.180)
- # [17:20] * Joins: shepazu (schepers@128.30.52.30)
- # [17:35] * tlr is now known as tlr-bbl
- # [17:41] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
- # [18:22] * Joins: Lachy (Lachlan@85.196.122.246)
- # [19:07] * Quits: gsnedders (gsnedders@86.136.52.180) (Quit: Killin' teh intarwebs)
- # [19:08] * Joins: gsnedders (gsnedders@86.136.52.180)
- # [20:31] * Quits: chaals (chaals@89.130.83.193) (Quit: chaals)
- # [20:35] <ArtB> shepazu, re DOM L3 Events spec, do you have a "delta doc" that describes the major changes/diff for L2?
- # [20:41] <shepazu> ArtB: no, I don't, but we shoudl make one
- # [20:41] * shepazu is at a conference right now
- # [20:50] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [20:53] * tlr-bbl is now known as tlr
- # [21:07] * tlr is now known as tlr-off
- # [21:17] * Joins: chaals (chaals@89.130.83.193)
- # [21:34] * Quits: tlr-off (tlr@128.30.52.30) (Quit: tlr-off)
- # [21:47] * Joins: shepazu (schepers@128.30.52.30)
- # [22:28] * Quits: ArtB (ce846302@128.30.52.43) (Quit: CGI:IRC)
- # [22:45] * Quits: heycam (cam@124.168.33.158) (Quit: bye)
- # [23:06] * Joins: heycam (cam@130.194.72.84)
- # [23:18] <marcos> gsnedders: hehe :)
- # [23:23] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [23:37] * Joins: aroben (aroben@17.244.16.92)
- # [23:49] * Quits: aroben (aroben@17.244.16.92) (Ping timeout)
- # [23:54] * Joins: aroben (aroben@17.244.16.92)
- # Session Close: Tue Feb 03 00:00:00 2009
The end :)